The other thing I wonder about is that Tim seems to have put in some code that depends upon xercesImpl.jar in the latest release. (Is that correct, Tim?)
Tim, could Aaron's problems be related to that change?
I don't recall doing it on purpose, but it may have been a while back. I can check.
The other strange thing is I don't see any reference to our code in this stack trace.
That's what surprised me. The other piece that stood out is that the instrumenting completed before the stack trace, and testing (and hence coverage work) didn't start until after the stack trace.
[echo] (hackyBuildUtils) Instrumenting bytecode with JBlanket.
[echo] (hackyBuild) Installing hackystat web app to tomcat.
[echo] (hackyBuild) OK - Installed application at context path
/hackystat
[echo] (hackyBuild) Deploying soap services.
[java] java.lang.NullPointerException
....
[echo] (hackyBuild) Starting build on March 8 2005 11:50:33.
[echo] (hackyBuildUtils) Invoking junit for hackyAnt.
[echo] (hackyBuildUtils) Running JUnit tests on module hackyAnt.
[echo] (hackyBuildUtils) Sending JUnit data to hackystat.Is that right, or am I misinterpreting the build output? I'm not real familiar with the hackyBuild stuff (your standard targets and their order), though I'm pretty versed in Ant in general.
--Tim
